A practical approach to mitigate

A few years back, we realised that Transaction Cost Analysis (TCA) was a sine qua non of efficient trading across all asset classes, especially Fixed-income. However, one pressing problem was the absence of post-negotiation public data for most European bonds, and the realisation that a reporting process for bond transactions at industry level would not be immediately implementable. We noticed, too, that, for equities (irrespective of their liquidity level), pre- and post-trade information was easily retrievable via different channels; although we were not always able to access brokers’ indications of interest (IOIs), we could at least see their ranking by volume of transactions over any given period of time. To solve this problem of the lack of information on bonds, in particular corporate credits, we resorted to our Order Management System (OMS) and its historical transaction database; however, even these data were not comprehensive enough for in-depth post-trade analysis.

With respect to Fixed Income, we thought first about making more use of the OMS post-negotiation data, combining them with other information (in particular the indicative bid-offer spreads captured at the time of the transactions) from our different multi-bank trading platforms and then integrating them all within a proprietary TCA-dedicated tool.

While this in-house application was under construction, and before thinking about analysing trades, transaction costs and broker/dealer performance, we first had to define similar order categories based on the instrument’s liquidity level. With this in mind, the challenge lay in the fact that it is more difficult to determine the (il)liquidity of a bond than it is that of an equity. There are many more liquidity criteria governing bonds than governing equities; these are related to the primary issue size, a bond’s duration, its rating, its capital structure, its category (sovereign bond, financial or corporate credit), the possible presence of credit derivatives and, finally, the order size. To obtain the most comprehensive analysis possible, we also wanted to input all our transactions into the application, not only those done across one or more trading platforms.

Once the TCA application was up and running, we could then carry out regular segmented analyses of the transaction costs and compare the execution performances of our broker/dealers. However, more recently, with the resurgence in volatility, the widening bid-offer spreads across trading platforms, the growing market fragmentation and the occasional reticence of certain dealers to take a position (at quarter’s end, for example), we upgraded the tool, incorporating into it a search engine connected to our historical transaction database.

In practice, before executing an order in the secondary market, the trader consults the banks’ axes and, in the case of less liquid instruments, uses this search engine (liquidity discovery tool), which, in a flash, lists the Top 6 banks for the underlying instrument as traded with Candriam over the desired period: a week, a month, a quarter, a half-year or a full year. Although the search is focused on the secondary market, it can, if necessary, include the
primary market. The search can also be expanded by adding the historical activity on other instruments from the same issuer. This could help detect investment opportunities on other, more liquid bonds and thus strengthen the dialogue with portfolio managers. Even if the search doesn’t prove wholly satisfying, we can still submit our ticket into an integrated order book at a price agreed with the manager.

The liquidity discovery tool enhances our counterparty selection process and our decision-taking prior to the transaction. Having such a search engine at one’s disposal improves the RFQ (request for a quote) process across the multi-bank trading platforms and helps target the counterparties most likely to conclude a deal at the best price. It also informs us further about our counterparties – which include regional banks – and their business growth. In an ever more fragmented bond market, this search engine, based on in-house post-trade data, actually helps improve the pre-trade function through its targeted approach. Obviously, the effectiveness of such a tool depends on the size and accuracy of the internal transactional database.

Not only can TCA shows that orders are executed in accordance with the in-house policy for best execution, it can also, in the case of the bond market, help the trader select his broker/dealers more carefully and find liquidity more efficiently. Regular TCA reports enhance relations with our counterparties; this is very important, especially when doing business over the telephone. Our pragmatic TCA approach to Fixed Income complements other initiatives in the pre-trade function currently being implemented by certain IT firms, e.g., simplified protocols for the transmission of banks’ axes, runs and inventory.

Direct access in the OMS – to a one-stop shop that incorporates real-time market data, easily accessible historical transactions and axes – would represent real progress for buy-side trading desks and would substantially increase the chances of executing the orders processed on the least liquid instruments. The more systematic use of search engines on transactions and on axes will pave the way to the development of smart order routers adapted to the particularities of the bond market.

Fabien Orève
Global Head of Trading